- telegram/bot.go: clientForProvider now returns NewClaudeClient() for
'claude' provider (was incorrectly falling back to DeepSeekClient which
uses OpenAI wire format, breaking Anthropic API calls)
- api/server.go: fix scan_interval_minutes schema default (3, not 60);
POST /api/strategies now clearly states config is OPTIONAL with complete
working defaults; POST /api/traders removes redundant GET workflow note
- telegram/agent/prompt.go: simplify strategy creation — just POST {name}
without config (backend applies full working defaults automatically);
only include config when user requests custom settings
Problem: callWithRequest/Full/Stream all called client.buildRequestBodyFromRequest
directly (not via hooks), so ClaudeClient could never override it. This meant
tool calling sent OpenAI format to Anthropic (wrong field names, wrong roles).
Changes:
mcp/interface.go
- Add buildRequestBodyFromRequest(*Request) map[string]any to clientHooks
- Improve comments: document what each hook group does and why
mcp/client.go
- All three paths (callWithRequest, callWithRequestFull, CallWithRequestStream)
now call client.hooks.buildRequestBodyFromRequest — ClaudeClient picks up
mcp/claude_client.go
- Full rewrite with format comparison table in package doc
- buildRequestBodyFromRequest: produces correct Anthropic wire format
* system prompt → top-level "system" field
* tools: parameters → input_schema, no "type:function" wrapper
* tool_choice "auto" → {"type":"auto"} object
* assistant tool calls → content[{type:tool_use, id, name, input}]
* role=tool results → role=user content[{type:tool_result,...}]
* consecutive tool results merged into single user turn
- convertMessagesToAnthropic: handles all three message types
- parseMCPResponseFull: extracts text + tool_use blocks
- parseMCPResponse: delegates to parseMCPResponseFull
All mcp and agent tests pass.
Migrate the Telegram bot agent from an XML tag hack (<api_call>) to
OpenAI-native function calling via CallWithRequestFull.
Key changes:
- mcp/interface.go: add parseMCPResponseFull to clientHooks interface
- mcp/client.go: route callWithRequestFull through hooks for overridability
- mcp/claude_client.go: override parseMCPResponseFull for Claude response
format (tool_use blocks instead of choices[].message.tool_calls)
- telegram/agent/agent.go: rewrite Run() to use CallWithRequestFull;
define api_request tool with JSON Schema; implement tool-call loop
with role="tool" result messages; remove XML parsing entirely
- telegram/agent/apicall.go: remove parseAPICall (dead code)
- telegram/agent/prompt.go: simplify — remove XML format instructions,
replace with concise api_request tool usage instructions
- telegram/agent/agent_test.go: rebuild all tests using LLMResponse
objects; add TestNarrationStructurallyImpossible, TestOnChunkCalledWithFinalReply,
TestToolCallIDPropagated; remove XML-specific tests
Architecture advantage: with native function calling, the LLM returns
EITHER ToolCalls OR Content — never both. Narration is now structurally
impossible at the protocol level, not just enforced by prompt rules.
All 11 agent tests pass. mcp package tests pass.
- Rewrite NO NARRATION rule: response is EITHER api_call tag alone OR
final text reply — no text before api_call under any circumstances
- Ban all narration patterns: 现在我将/好的/正在/I will/Let me etc.
- Add 'create strategy + create trader + start' full setup workflow
- Add 12 automated tests covering:
- No narration leaking to user (5 narration variants tested)
- api_call tag never leaks to user
- Full setup workflow: POST strategy → verify → POST trader → start
- Start existing trader workflow
- Max iterations safety, tag stripping, parser edge cases
- Add Telegram bot with long-polling and AI agent loop (api_call tool)
- SSE streaming with real-time message editing and ⏳ placeholder
- Account state injection at conversation start (models, exchanges,
strategies, traders, per-trader PnL and statistics)
- Lane semaphore per chat serializes concurrent messages (60s timeout)
- Idle timeout watchdog (60s) prevents hung streaming connections
- Look-ahead buffer prevents partial <api_call> tag leaking to user
- Fix PUT /strategies/:id to merge config (read-then-merge pattern)
- Add route registry with full API schema for LLM documentation
- Add TelegramConfig store and Web UI config modal
- Add GetAnyEnabled to AIModel store for bot LLM client selection
- Add 'id' to Language type in translations.ts
- Add ~1000 Indonesian translation keys covering all UI sections
- Update LanguageContext to persist 'id' in localStorage
- Add ID button to Header.tsx language toggle
- Add �� option to HeaderBar.tsx desktop dropdown and mobile toggle
- Add Indonesian translations to inline text objects in
LoginRequiredOverlay, StrategyMarketPage, PositionHistory
Closes #XX
This PR adds support for Hyperliquid's Unified Account mode where Spot USDC
balance can be used as collateral for Perpetual trading.
Changes:
- Add HyperliquidUnifiedAcct field to Exchange config (default: true)
- Update HyperliquidTrader to support unified account mode
- When enabled, Spot USDC balance is added to available trading balance
- Update API request/response structs for unified account toggle
- Update trader config propagation from exchange config
This aligns with Hyperliquid's roadmap to make Unified Account the default.
Add two new coin source options for Hyperliquid trading:
- hyper_all: All available Hyperliquid perpetual coins (229 coins)
- hyper_main: Top N coins by 24h volume (default 20)
Changes:
- Add CoinSourceConfig fields: UseHyperAll, UseHyperMain, HyperMainLimit
- Add provider/hyperliquid/coins.go with caching (24h) and volume-based sorting
- Add source types 'hyper_all' and 'hyper_main' to GetCandidateCoins()
- Support mixing with other sources in 'mixed' mode
- Add source tag formatting for UI display
This ensures traders using Hyperliquid can select coins that are actually
available on the exchange, avoiding 'symbol not found' errors.
- Moved from project root to docs/community/ for better organization
- Added Telegram community channel to official accounts list
- Made all links clickable with proper markdown formatting
- Added direct GitHub issue link for impersonation reports
- Added navigation links and reference from community README
- Moved from project root to docs/community/ for better organization
- Added Telegram community channel to official accounts list
- Made all links clickable with proper markdown formatting
- Added direct GitHub issue link for impersonation reports
- Added navigation links and reference from community README
* feat(i18n): add 42 translation keys for TraderConfigModal
- Add new translation keys for all hardcoded Chinese strings
- Replace hardcoded UI text with t('key', language) calls
- Support both English and Chinese languages
Modified files:
- web/src/i18n/translations.ts: +88 lines (42 new keys)
- web/src/components/TraderConfigModal.tsx: replaced 48 hardcoded strings
* feat(i18n): add consolidated translation keys (en + zh + es)
- 275+ translation keys from 8 strategy components
- 3 languages: English, Chinese, Spanish
- Ready for integration into translations.ts
- Pre-aggregated exports for zhStrategy, enStrategy, esStrategy
Related to PR #1343 (maker95) and #1374 (xsa-dev)
- Add new translation keys for all hardcoded Chinese strings
- Replace hardcoded UI text with t('key', language) calls
- Support both English and Chinese languages
Modified files:
- web/src/i18n/translations.ts: +88 lines (42 new keys)
- web/src/components/TraderConfigModal.tsx: replaced 48 hardcoded strings
- Rename 'Bias Ratio' to 'Bias Strength' (偏向强度)
- Add direction modes explanation (neutral/long/short/long_bias/short_bias)
- Show actual buy/sell ratios for both long_bias and short_bias modes
- Add bilingual support (Chinese/English)
- Clarify that X% applies differently to long_bias vs short_bias
- Add adaptivePriceRound() in store/position.go for database storage
- Update position_builder.go to use adaptive precision for entry/exit prices
- Add Gate to OrderSync skip list in auto_trader.go
- Add debug logging in gate/order_sync.go for price parsing issues
- Create web/src/utils/format.ts with formatPrice() for frontend display
- Update TraderDashboardPage.tsx and PositionHistory.tsx to use adaptive formatting
Fixes issue where meme coin prices (e.g. 0.000000166) displayed as 0.0000
- Add enable_direction_adjust and direction_bias_ratio to GridStrategyConfig
- Add Direction Auto-Adjust section in GridConfigEditor
- Include toggle switch, bias ratio slider, and explanation text
- Support both Chinese and English translations
- Add server time synchronization for KuCoin API to fix timestamp error (400002)
- Return empty list instead of error when no available coins (ai500.go)
- Save cycle record even when no candidate coins (show in frontend without red error)
- Update Claude icon to Anthropic dark brand color (#141413)
- Add exchange and AI model icons to README.md and README.ja.md
- Add KuCoin order sync with proper API response parsing
- Use openFeePay/closeFeePay to determine open/close trades
- Get contract multiplier from API for accurate qty calculation
- Fix price rounding: 2 decimals -> 8 decimals for low-price coins
- Add comprehensive tests for trades, positions, and P&L
Gate.io Integration:
- Add Gate trader with full Trader interface implementation
- Add order_sync.go for background trade synchronization
- Fix quantity display (convert contracts to actual tokens via quanto_multiplier)
- Fix fill price return in OpenLong/OpenShort/CloseLong/CloseShort
- Add Gate-specific CoinAnk K-line data source support
- Add Gate to supported exchanges in frontend and backend
- Add Gate/KuCoin logo SVG icons
Trader Package Refactoring:
- Move exchange-specific code into subdirectories (binance/, bybit/, okx/, bitget/, hyperliquid/, aster/, lighter/, gate/)
- Create types/ package for shared types to avoid circular dependencies
- Move TraderTestSuite to trader/testutil package to avoid import cycles
- Update market.GetWithExchange to support exchange-specific data
- Support both wrapped (fillList) and direct array response formats
- Fix tradeSide parsing for one-way mode (buy_single/sell_single)
- Fix fee extraction from nested feeDetail array structure
- Add oi_low as independent source_type for short opportunities
- Redesign Mixed mode with card-based selector (2x2 grid)
- Show combination summary with total coin limit
- Support both Chinese and English languages
- Change default limits to 10 for OI Top and OI Low
